I wonder if the augmented reality tech that Yelp is using on the iphone could be overlaid on a Justin.tv stream. Anyone watching a stream could click on a business in the video and connect to their site|Opentable|Yelp. If we could get solid object recognition working we could link to Amazon product pages (revenue!!) or such. Find a bottle of Pepsi|Coke in a stream, click on it, get a coupon for .50 off your next Whopper meal with a large soda. (revenue??)
I'm pretty sure this is technically possible. Eric and I both worked hard to get around Flash's often insane security stuff and make sure the bitmap of the video stream is available if you want it. So you should be able to just grab the bitmap periodically, do some computation on it, and then overlay whatever you want in a Flash layer that sits on top of our video layer.
I think jtv augmented reality apps would be great!
-Traffic monitoring. Partner with a few delivery companies, and you can start to push live traffic feeds to the web. Start in one city to test feasibility, then grow.
-Maps. Let people register to record a trip from their home to work. Send the package, they set it up on their dashboard, record the round-trip, then send the package to the next location.
-Secret shopping. Boy, do they treat people poorly at some places. Please take a live feed, so employees can be fired or rewarded on the spot.
- Recording Clubs live, so I know what it looks like before I go and spend 10 bucks to get in and find it empty.
We have http://justin.tv/dnalounge - that one has worked out pretty well, but obviously jwz isn't your average club owner. I'd love to see someone start a jtv-based business that helps all the other clubs get online (but I have no idea what YC would think of it - just in case that needs to be said!).
The traffic monitoring idea could work on its own. UPS & Fedex already tracks their trucks (I think). If you could buy access to that data you could publish detailed drive time info for a big part of the developed world. The data from FedEx|UPS wouldn't have any privacy issues around it; easy for them to implement a feed, no really advanced new tech involved anywhere in the process. I think the majority of the work would be licensing/business relationship stuff.
I'm not sure who you could sell the data to? Set yourself up to be acquired by Google/Navtech/XM Sirus? Something to think about.
Let's say you partner with a few pizza and taco delivery companies in San Fransisco. They record the traffic on the go, and the feed gets pushed to the website. It could be useful to have a snapshot of traffic before leaving home/office. I think it is easier to cut a deal with small shops, than big Fedex.
Forget traffic, what if we simply wanted a safer world, and decided to record what's happening in the streets while driving. This live feed can make the world such a better place.
A smaller business probably wouldn't have GPS tracking already implemented so our startup would have to be involved with the capitol cost of equipping their vehicles with that hardware at $400-$500/vehicle plus the apps/infrastructure to aggregate and upload the data to us.
Also I think the coverage would be spotty/sporadic. Who orders take-out in the morning? There goes the morning commute customer base. FedEx, UPS have trucks on the road 24/7. Also the time involved in the bus dev working with 200-500 pizza|taco|Chinese restaurants would be huge.
Three data sources (FedEx, UPS, DHL?), a few tens of millions of consumers accessing the data via two or three methods is doable. I just can't figure out how to make make any money on it.
Agree with gojomo that cell-tower tracking is more efficient. Besides, why use delivery vehicles? Buses would probably tell you just as much about traffic conditions and that info is already up there for free.
Cell towers already have this info -- and the volume should make up for the imprecision, as by now an overwhelming majority of vehicles in motion must have an active cell phone inside. So video is overkill for traffic monitoring.
From a practical POV I don't think you could ever, ever, ever get around the perceived privacy issue of using cell phones. (Maybe even a real concern.)
As to the using the bus data I have run some tests using the SF Muni dataset (nextbus.com). The speed of the buses in the dense parts of the city doesn't correlate with the speed of traffic. They stop too often and the timing of the stops is inconsistent. Here in SF the buses often move over a fixed distance (20 blocks, Van Ness) in the same amount of time at 2AM and 5PM.
Dash is a great idea (we own two!) but obviously the money hasn't worked out.
It's not much of a business idea, but I've recently been thinking about a way for communities (esp in high crime areas) to do an online neighborhood watch. IE, neighbors put webcams in their front windows and the site presents combined (customizable) views of these webcams a la security camera displays. What if Grandma could email the police a security camera sequence showing who vandalized the library? You could even imagine communities providing bounties for any tips leading to a conviction, so kids in India can protect Walmart parking lots for us.
There's an effort to use Justin.tv as a neighborhood security watch at Ourblock.tv, which is the successor to Adam's Block (if you remember that news report a while back) However, the team coding the site all have full time jobs, and development is moving slowly (but steadily).
My web presence is getting drawn more and more to live video, and my current solution works...sort of, and I'm always looking for better alternatives-- but it's good to see other alternatives that aren't just out of the box.
I know the justin.tv guys are on here, would love to ping some ideas off you and make some connections.
